Part1. 繪製地圖安裝繪製地圖所需程式

Part1.繪製地圖安裝繪製地圖所需程式

ssc install spmap ssc install shp2dtassc install mif2dta

讀取 CHN_adm1地圖檔。創出chinaprovince以及 coord兩個

檔案shp2dta using CHN_adm1, database (chinaprovince) coordinates(coord) genid(id) gencentroids(c)

*打開 chinaprovince檔案，刪除 Paracel Islands

把想顯示於地圖的 data放進 chinaprovince中

繪出 2007年中國 FDI分布圖spmap FDI_2007 using coord, id(id) clnumber(5)

可點選 start graph editor編輯

Part2.建立空間計量矩陣findit spmat

點選第一個超連結

點選 (click here to install)進行下載

讀取 chinaprovince，並將 x_c 跟y_c改名

改為 longitude 與 latitude

生成距離矩陣spmat idistance idistance_province longitude latitude, id(id) dfunction(euclidean) normalize(row)

生成 stata可讀之檔案idistance_province.spmat

spmat save idistance_province using idistance_province.spmat

(此時 stata資料夾出現idistance_province.spmat)

將剛剛的idistance_province.spmat檔案轉為我們可閱讀的 txt檔spmat export idistance_province using idistance_province.txt

此時 stata資料夾出現idistance_province.txt

生成相鄰矩陣spmat contiguity contiguity_province using coord, id(id) normalize(row)

生成 stata可讀之檔案spmat save contiguity_province using contiguity_province.spmat

將剛剛的contiguity_province.spmat檔案轉為我們可閱讀的 txt檔spmat export contiguity_province using contiguity_province.txt

此時 stata資料夾出現contiguity_province.txt

Part.3 跑空間計量findit xsmle

點選第一個超連結

往下拉，點選 (click here to install)進行下載

spmat use W using "idistance_province.spmat "(使用距離矩陣跑空間計量 )spmat use W using "contiguity_province.spmat"(使用相鄰矩陣跑空間計量 )

告知 stata要跑 panel data，才可以跑空間計量

xtset id year

開啟資料 data，公布年度 _上課用

xsmle FDI FT GDP WAGE OPEN CFDI , wmat(W) mode(sdm) re(其中第一個變數 stata會自動設定為 Y ，之後都是X ， re為隨機效果模型， fe為固定效果模型。mode可選 sar 與 sdm)

CFDI .0927104 .0070784 13.10 0.000 .078837 .1065837

OPEN 52617.83 15357.92 3.43 0.001 22516.87 82718.8

WAGE -.8204718 .4860468 -1.69 0.091 -1.773106 .1321623

GDP .6832036 .5361676 1.27 0.203 -.3676656 1.734073

FT 94.43746 84.33592 1.12 0.263 -70.85791 259.7328

Direct

sigma_e 1.43e+07 2820431 5.08 0.000 8799255 1.99e+07

lgt_theta -2.983231 .2074382 -14.38 0.000 -3.389802 -2.57666

Variance

rho .1448937 .3745166 0.39 0.699 -.5891454 .8789328

Spatial

CFDI -.0250511 .059445 -0.42 0.673 -.1415612 .091459

OPEN 7278.992 46083.16 0.16 0.874 -83042.34 97600.32

WAGE -1.156691 1.337933 -0.86 0.387 -3.778992 1.465609

GDP 2.992128 3.799625 0.79 0.431 -4.455 10.43926

FT 166.7804 605.3808 0.28 0.783 -1019.744 1353.305

Wx

_cons 34602.23 25657.5 1.35 0.177 -15685.55 84890.02

CFDI .0925311 .0064227 14.41 0.000 .0799428 .1051194

OPEN 52942.06 14912.52 3.55 0.000 23714.05 82170.06

WAGE -.8209484 .457982 -1.79 0.073 -1.718577 .0766798

GDP .5898734 .4588287 1.29 0.199 -.3094144 1.489161

FT 95.36754 91.50272 1.04 0.297 -83.9745 274.7096

Main

FDI Coef. Std. Err. z P>|z| [95% Conf. Interval]

Log-likelihood = -992.2360

CFDI .0884814 .0062748 14.10 0.000 .076183 .1007798

OPEN 31394.1 12006.58 2.61 0.009 7861.647 54926.56

WAGE -.6336338 .4152306 -1.53 0.127 -1.447471 .1802032

GDP .4246668 .5147828 0.82 0.409 -.5842889 1.433623

FT 69.75591 71.38884 0.98 0.329 -70.16364 209.6755

Direct

sigma2_e 8724978 1279760 6.82 0.000 6216695 1.12e+07

Variance

rho -.0728405 .4367999 -0.17 0.868 -.9289525 .7832716

Spatial

CFDI -.0074108 .0547492 -0.14 0.892 -.1147173 .0998957

OPEN 11053.53 40357.06 0.27 0.784 -68044.86 90151.92

WAGE -1.446171 1.164901 -1.24 0.214 -3.729335 .8369927

GDP 3.910964 3.290304 1.19 0.235 -2.537914 10.35984

FT -37.54984 494.9329 -0.08 0.940 -1007.6 932.5008

Wx

CFDI .0881605 .0050674 17.40 0.000 .0782286 .0980925

OPEN 31436 12066.89 2.61 0.009 7785.338 55086.66

WAGE -.6528307 .388018 -1.68 0.092 -1.413332 .1076707

GDP .3990044 .3815596 1.05 0.296 -.3488387 1.146847

FT 68.49903 73.67781 0.93 0.353 -75.90682 212.9049

Main



Sem模型xsmle FDI FT GDP WAGE OPEN CFDI , ematrix(W) mode(sem) re

將 W 改為相鄰矩陣spmat use W using "contiguity_province.spmat" ,replace

重複剛剛的步驟xsmle FDI FT GDP WAGE OPEN CFDI , wmat(W) mode(sdm) re

WAGE -.8280938 .2907724 -2.85 0.004 -1.397997 -.2581904

GDP .4329611 .4719747 0.92 0.359 -.4920923 1.358015

FT 115.4091 73.59384 1.57 0.117 -28.83218 259.6504

Direct

sigma_e 1.19e+07 2355401 5.05 0.000 7280572 1.65e+07

lgt_theta -2.959432 .2054211 -14.41 0.000 -3.36205 -2.556814

Variance

rho .3649768 .1315159 2.78 0.006 .1072104 .6227432

Spatial

CFDI -.0460164 .0178403 -2.58 0.010 -.0809827 -.0110501

OPEN 1975.193 24510.05 0.08 0.936 -46063.61 50014

WAGE -.2951333 .4505019 -0.66 0.512 -1.178101 .5878342

GDP 1.385675 .9683436 1.43 0.152 -.512244 3.283593

FT 315.4662 163.4135 1.93 0.054 -4.81836 635.7507

Wx

_cons 20322.98 9976.056 2.04 0.042 770.2701 39875.69

CFDI .0953237 .0064428 14.80 0.000 .082696 .1079513

OPEN 51826.03 13089.78 3.96 0.000 26170.53 77481.54

WAGE -.7870454 .2761907 -2.85 0.004 -1.328369 -.2457217

GDP .2640631 .4217304 0.63 0.531 -.5625133 1.09064

FT 83.52678 79.32219 1.05 0.292 -71.94186 238.9954

Main



CFDI .0892304 .0050846 17.55 0.000 .0792647 .0991961

OPEN 33359.6 11465.38 2.91 0.004 10887.86 55831.34

WAGE -.5767623 .2488954 -2.32 0.020 -1.064588 -.0889363

GDP .3193532 .4195972 0.76 0.447 -.5030423 1.141749

FT 93.11735 58.38873 1.59 0.111 -21.32246 207.5572

Direct

sigma2_e 7293312 1084031 6.73 0.000 5168651 9417974

Variance

rho .3369109 .1406159 2.40 0.017 .0613088 .612513

Spatial

CFDI -.0385547 .0164329 -2.35 0.019 -.0707626 -.0063468

OPEN -12976.88 21164.02 -0.61 0.540 -54457.6 28503.85

WAGE -.333219 .3885812 -0.86 0.391 -1.094824 .4283861

GDP 1.470184 .8387266 1.75 0.080 -.1736897 3.114058

FT 239.7663 130.445 1.84 0.066 -15.90124 495.4339

Wx

CFDI .0895429 .0050829 17.62 0.000 .0795806 .0995051

OPEN 33836.74 10791.88 3.14 0.002 12685.05 54988.42

WAGE -.5378699 .2190107 -2.46 0.014 -.967123 -.1086169

GDP .1562309 .3554739 0.44 0.660 -.5404851 .8529469

FT 66.30069 63.39159 1.05 0.296 -57.94453 190.5459

Main



檢定使用 sdm模型或是 sar模型

如果拒絕 H0，則使用 sdm

test [Wx]FT=[Wx]GDP=[Wx]WAGE=[Wx]OPEN=[Wx]CFDI=0

拒絕 H0，使用 sdm

Prob > chi2 = 0.0186

chi2( 5) = 13.57

( 5) [Wx]FT = 0

( 4) [Wx]FT - [Wx]CFDI = 0

( 3) [Wx]FT - [Wx]OPEN = 0

( 2) [Wx]FT - [Wx]WAGE = 0

( 1) [Wx]FT - [Wx]GDP = 0

. test [Wx]FT=[Wx]GDP=[Wx]WAGE=[Wx]OPEN=[Wx]CFDI=0

檢定使用 sdm模型或是 sem模型

如果拒絕 H0，則使用 sdmtestnl ([Wx]FT = -[Spatial]rho*[Main]FT) ([Wx]GDP = -[Spatial]rho*[Main]GDP) ([Wx]WAGE = -[Spatial]rho*[Main]WAGE) ([Wx]OPEN = -[Spatial]rho*[Main]OPEN) ([Wx]CFDI = -[Spatial]rho*[Main]CFDI)

拒絕 H0，使用 sdm

Prob > chi2 = 0.0431

chi2(5) = 11.45

(5) [Wx]CFDI = -[Spatial]rho*[Main]CFDI

(4) [Wx]OPEN = -[Spatial]rho*[Main]OPEN

(3) [Wx]WAGE = -[Spatial]rho*[Main]WAGE

(2) [Wx]GDP = -[Spatial]rho*[Main]GDP

(1) [Wx]FT = -[Spatial]rho*[Main]FT

> OPEN = -[Spatial]rho*[Main]OPEN) ([Wx]CFDI = -[Spatial]rho*[Main]CFDI)

. testnl ([Wx]FT = -[Spatial]rho*[Main]FT) ([Wx]GDP = -[Spatial]rho*[Main]GDP) ([Wx]WAGE = -[Spatial]rho*[Main]WAGE) ([Wx]

使用 hausmam test檢定固定效果模型或隨機效果模型

先跑隨機效果模型，並儲存結果

xsmle FDI FT GDP WAGE OPEN CFDI , wmat(W) mode(sdm) re

estimates store re

再先跑固定效果模型，並且儲存結果

xsmle FDI FT GDP WAGE OPEN CFDI , wmat(W) mode(sdm) fe

estimates store fe

檢定使用固定效果模型或隨機效果模型

hausman re fe

拒絕 H0，使用固定效果模型

.

Prob>chi2 = 0.0523

= 5.90

chi2(2) = (b-B)'[(V_b-V_B)^(-1)](b-B)

Test: Ho: difference in coefficients not systematic

B = inconsistent under Ha, efficient under Ho; obtained from xsmle

b = consistent under Ho and Ha; obtained from xsmle

CFDI .0953237 .0895429 .0057808 .003959

OPEN 51826.03 33836.74 17989.3 7407.956

WAGE -.7870454 -.5378699 -.2491755 .1682724

GDP .2640631 .1562309 .1078323 .2269248

FT 83.52678 66.30069 17.22609 47.68141

re fe Difference S.E.

(b) (B) (b-B) sqrt(diag(V_b-V_B))

Coefficients

Part4. Moran’s I

findit spatwmat

點選第二個連結


讀取 chinaprovince

spatwmat, name(Chinaweights) xcoord(longitude) ycoord(latitude) band(0 8)

You are advised to extend the distance band

Beware! 2 locations have no neighbors

Smallest maximum distance: 24.06

Largest minimum distance: 9.86

Maximum distance: 43.1

3rd quartile distance: 19.4

Median distance: 13.2

1st quartile distance: 8.6

Minimum distance: 0.7

Friction parameter: 1

Distance band: 0 < d <= 8

Dimension: 31x31

1. Inverse distance weights matrix Chinaweights

The following matrix has been created:

. spatwmat, name(Chinaweights) xcoord(longitude) ycoord(latitude) band(0 8)

下載繪製Moran’s I工具findit splagvar

點選第一個連結


讀取檔案moran_FDI

splagvar FDI_2007, wname(Chinaweights) wfrom(Stata) moran(FDI_2007) plot(FDI_2007)

Part1. 繪製地圖 安裝繪製地圖所需程式

Documents

Part1. 繪製地圖安裝繪製地圖所需程式